In a server, computing resources and storage resources are key factors of entire system performance. With rapid development of microelectronic technologies, performance of a processor is being multiplied. Although processor technologies are progressing, a memory capacity is still limited. Due to the memory capacity limitation, a utilization rate of the server cannot increase with improvement of processor performance. Reasons are as follows.
First, currently the processor performance is improved primarily using the number of cores and a cache rather than a clock speed; secondly, the number of cores, the cache capacity, and the number of threads increase with a decrease of an operating frequency of a central processing unit (CPU), and a deficient increase of the memory capacity leads to imbalance; and thirdly, 64-bit applications and operating systems require a larger memory.
Seen from a perspective of a utilization rate of an entire server, the utilization rate of the server is no longer restricted by performance of the CPU, but primarily restricted by the memory capacity and a read/write rate of memory data.
Currently, the read/write rate of the memory data is restricted by manufacturing industrial technologies, and a read/write speed of a storage can hardly be increased. Therefore, only how to expand the memory capacity can be considered. A technical issue that currently needs to be solved is how to expand the memory capacity and develop a memory expansion device applicable to a plurality of interfaces and a plurality of configurations and switchable to storage expansion hardware so as to enhance server performance.